Abstract⎯This paper introduces a continuous-time, controllable Markov process model of a power-managed system. The system model is composed of the corresponding stochastic models of the service queue and the service provider. The system environment is modeled by a stochastic service request process. The problem of dynamic power management in such a system is formulated as a policy optimization problem and solved using an efficient “policy iteration ” algorithm. Compared to previous work on dynamic power management, our formulation allows better modeling of the various system components, the power-managed system as a whole, and its environment. In addition it captures dependencies between the service queue and service provider status. Finally, the resulting power management policy is asynchronous, hence it is more power-efficient and more useful in practice. Experimental results demonstrate the effectiveness of our policy optimization algorithm compared to a number of heuristic (time-out and N-policy) algorithms. I.

Redundancy based on a parity encoding has been proposed for insuring that disk arrays provide highly reliable data. Parity-based redundancy will tolerate many independent and dependent disk failures (shared support hardware) without on-line spare disks and many more such failures with on-line spare disks. This paper explores the design of reliable, redundant disk arrays. In the context of a 70 disk strawman array, it presents and applies analytic and simulation models for the time until data is lost. It shows how to balance requirements for high data reliability against the overhead cost of redundant data, on-line spares, and on-site repair personnel in terms of an array’s architecture, its component reliabilities, and its repair policies.

The goal of a dynamic power management policy is to reduce the power consumption of an electronic system by putting system components into different states, each representing certain performance and power consumption level. The policy determines the type and timing of these transitions based on the system history, workload and performance constraints. In this paper, we propose a new abstract model of a power-managed electronic system. We formulate the problem of system-level power management as a controlled optimization problem based on the theories of continuous-time Markov decision processes and stochastic networks. This problem is solved exactly and efficiently using a "policy iteration" approach. Our method is compared with existing heuristic approaches for different workload statistics. Experimental results show that power management method based on Markov decision process outperforms heuristic approaches in terms of power dissipation savings for a given level of system performanc...

In this paper, we introduce a new technique for modeling and solving the dynamic power management (DPM) problem for systems with complex behavioral characteristics such as concurrency, synchronization, mutual exclusion and conflict. We model a power-managed distributed computing system as a controllable Generalized Stochastic Petri Net (GSPN) with cost. The obtained GSPN model is automatically converted to an equivalent continuous-time Markov decision process. Given the delay constraints, the optimal power management policy for system components as well as the optimal dispatch policy for requests are calculated by solving a linear programming problem based on the Markov decision process. Experimental results show that the proposed technique can achieve more than 20% power saving compared to other existing DPM techniques.

With the increasing complexity of multiprocessor and distributed processing systems, the need to develop efficient and accurate modeling methods is evident. Fault-tolerance and degradable performance of such systems has given rise to considerable interest in models for the combined evaluation of performance and reliability [1, 2]. Most of these models are based upon Markov or semi-Markov reward processes. Beaudry [1] proposed a simple method for computing the distribution of performability in a Markov reward process. We present two extensions of Beaudry's approach. First, we generalize the method to a semi-Markov reward process. Second, we remove the restriction requiring the association of zero reward to absorbing states only. Such reward models can be used to evaluate the effectiveness of degradable fault-tolerant systems. We illustrate the use of the approach with three interesting applications.

Abstract--Three methods for numerical transient analysis of Markov chains, the modified Jensen's method (Jensen's method with steady-state detection of the underlying DTMC and computation of Poisson probabilities using the method of Fox and Glynn [1]), a third-order L-stable implicit Runge-Kutta method, and a second-order L-stable method, TR-BDF2, are compared. These methods are evaluated on the basis of their performance (accuracy of the solution and computational cost) on stiff Markov chains. Steady-state detection in Jensen's method results in large savings of computation time for Markov chains when mission time extends beyond the steady-state point. For stiff models, computation of Poisson probabilities using traditional methods runs into underflow problems. Fox and Glynn's method for computing Poisson probabilities avoids underflow problems for all practical problems and yields highly accurate solutions. We conclude that for mildly stiff Markov chains, the modified Jensen's method is the method of choice. For stiff Markov chains, we recommend the use of the L-stable ODE methods. If low accuracy (upto eight decimal places) is acceptable, then TR-BDF2 method should be used. If higher accuracy is desired, then we recommend third-order implicit Runge-Kutta method. 1.

A dummy traffic strategy is described that can be implemented by mix nodes in an anonymous communication network to detect and counter active (n-1) attacks and their variants. Heartbeat messages are sent anonymously from the mix node back to itself in order to establish its state of connectivity with the rest of the network. In case the mix is under attack, the flow of heartbeat messages is interrupted and the mix takes measures to preserve the quality of the anonymity it provides by introducing decoy messages.

We explain how to apply statistical techniques to solve several language-recognition problems that arise in cryptanalysis and other domains. Language recognition is important in cryptanalysis because, among other applications, an exhaustive key search of any cryptosystem from ciphertext alone requires a test that recognizes valid plaintext. Written for cryptanalysts, this guide should also be helpful to others as an introduction to statistical inference on Markov chains. Modeling language as a finite stationary Markov process, we adapt a statistical model of pattern recognition to language recognition. Within this framework we consider four welldefined language-recognition problems: 1) recognizing a known language, 2) distinguishing a known language from uniform noise, 3) distinguishing unknown 0th-order noise from unknown 1st-order language, and 4) detecting non-uniform unknown language. For the second problem we give a most powerful test based on the Neyman-Pearson Lemma. For the oth...

We consider the problem of how to construct and maintain an overlay structured P2P network based on the small world paradigm. Two main attractive properties of a small world network are (1) low average hop distance between any two randomly chosen nodes, and (2) high clustering coefficient of nodes. Having a low average hop distance implies a low latency for object lookup, while having a high clustering coefficient implies the underlying network can effectively provide object lookup even under heavy demands (for example, in a flash crowd scenario). In this paper, we present a small world overlay protocol (SWOP) for constructing a small world overlay P2P network. We compare the performance of our system with that of other structured P2P networks such as Chord. We show that the SWOP protocol can achieve improved object lookup performance over the existing protocols. We also exploit the high clustering coefficient of a SWOP network to design an object replication algorithm that can effectively handle heavy object lookup traffic. As a result, a SWOP network can quickly and efficiently deliver popular and dynamic objects to a large number of requesting nodes. To the best of our knowledge, ours is the first piece of work that addresses how to handle dynamic flash crowds in a structured P2P network environment.

This paper presents the construction of a multicast service, called agreed multicast, that guarantees that messages arrive reliably and in the same total-order to all their destinations. ToTo, a novel family of protocols, implement the agreed multicast service of Transis, a communication sub-system for the High Availability project, currently developed at the Hebrew University of Jerusalem. This service is desired in distributed systems, and supports high level coordination among groups of processes in distributed applications. The ToTo protocols are genuinely symmetric. They are fairly simple and do not bear a significant processing burden. The cost of symmetric total ordering protocols can be measured by the number of messages from different machines needed to agree on the next set of messages to be delivered in total order. The ToTo protocols provide early delivery latency, and require as little as n 2 messages for delivery. Thus, we optimize the time that elapses from the point ...